ISO 14083 and the GLEC Framework: how freight transport emissions are calculated

The GHG Protocol splits third-party transport across two scope 3 categories, and the criterion separating them is not geographic. Category 4 covers freight paid for by the reporting company: inbound materials, movements between its own sites, and also the outbound movement of sold product whenever the transport invoice is its own. Category 9 covers the movement of already-sold product when someone else pays for it, plus the storage and retail that follow.

One consequence catches most people out the first time: outbound freight, which physically happens after the sale, is reported in a category labelled «upstream». The calculation guidance says it plainly — it is a purchased service, and purchased services sit upstream. This is not an inconsistency to be tidied up in the spreadsheet: it is the rule.

For an exporter that second question has an administrative answer, and it sits in the contract: the incoterm decides the category. Sold FOB, the buyer contracts and pays for the ocean leg, and that leg is category 9. Sold CIF or CFR, the same vessel, the same route and the same tonne are category 4. Changing sales terms from one season to the next moves emissions between categories with nothing having changed in the physical world, and the report has to explain that before anyone asks.

ISO 14083 was published in 2023 under a title that draws its own limits well: quantification and reporting of greenhouse gas emissions arising from transport chain operations. It is not a factor database, it is the method. It defines how a chain is cut up, which processes are included, how an intensity is assigned to a shipment and what has to be declared about the calculation.

The GLEC Framework, from the Smart Freight Centre, is the practical implementation of that method: version 3 incorporated the standard's provisions and, unlike ISO, it downloads free of charge. The GHG Protocol recognises it, CDP accepts it as a mechanism for calculating logistics emissions and SBTi uses it for target setting. In practice they work together — the standard says what to do and GLEC supplies the tables to do it with.

The standard does not talk about scopes

ISO 14083 avoids the split into scopes 1, 2 and 3 — it treats the distinction as commercially driven — and works with direct and indirect emissions instead. For a corporate inventory that means mapping the result afterwards: the standard hands you the leg's number, not the box it goes in.

It covers five categories, not fifteen

GLEC states how far it reaches inside scope 3: categories 1, 3, 4, 9 and 12. Business travel and employee commuting are explicitly out, because they move people rather than freight. Those two need a different factor source.

Conformance is not self-awarded

Saying a calculation follows the GLEC Framework is accurate and verifiable. Calling it «GLEC-conformant» is a different claim: it requires going through the Smart Freight Centre's assurance scheme.

The unit of calculation is neither the journey nor the shipment: it is the leg. The standard calls it a transport chain element and defines it as freight carried by a single vehicle or passing through a single hub. Every change of vehicle and every transfer point opens a new leg, calculated separately and only then added up.

The half that usually goes missing is the second one. A warehouse, a port terminal or a distribution centre are legs in their own right even though the freight moves no distance: they are measured by the tonnes leaving the centre, with zero distance, and their intensity is expressed per tonne handled rather than per tonne-kilometre. An inventory that only counts kilometres leaves out all the energy of refrigerated handling, which for fruit, fish or meat is not a rounding detail.

One more detail from the same chapter changes the denominator: the declared mass is that of the freight, not of the transport packaging. Pallets and containers do not add tonnes — except when what is being moved is the empty container, in which case the container is the freight.

A transport factor can measure two different things under the same name. One boundary covers vehicle operation, what burns while the vehicle moves. The other adds what it took to produce that fuel and bring it to the pump. ISO 14083 and GLEC work with the full boundary, and call that second part energy provision emissions.

The GHG Protocol sets a lower floor for categories 4 and 9: the mandatory minimum is the carrier's scope 1 and 2 emissions. Both boundaries comply, and the difference lies in how far above the floor you go. In GLEC's default values energy provision accounts for roughly a fifth of the total for road, and somewhat less for sea and air: reporting the minimum lowers the transport number by about that much. Which is why the boundary gets declared. Without stating it, two equally correct inventories cannot be compared.

The scope 3 calculation guidance sets out three routes, and the right one is the best method the available data supports.

Fuel-based

Start from the fuel the carrier consumed and apply its factor. It is the most accurate for CO₂, because the relationship between litres burned and emissions is direct. It works well when the vehicle carries a single customer's freight; on shared loads you have to allocate, and the standard sets the limiting criterion by mode: mass for road, rail and air, volume for sea.

Distance-based

Mass, distance and mode of each shipment, multiplied by a tonne-kilometre factor. It is the realistic method for most contracted freight, and the only feasible one downstream, where carrier fuel data never reaches you.

Spend-based

Money spent on the service times an economic factor. Useful for sizing the problem and deciding where better data is worth chasing; not useful for showing a reduction, because the number also falls when the freight rate does.

GLEC orders the same thing by data quality rather than by formula: primary, meaning measured; modelled, meaning a model fed with real parameters of the operation; and default, meaning the table. And one design decision behind those tables helps in reading them: the default values are deliberately set on the conservative side, so that improving your data is not penalised with a higher number. If replacing the table with carrier data pushes the result up sharply, the first thing to check is the vehicle class that had been selected.

A truck with a refrigeration unit does not consume what a dry one does, and that difference rarely shows up in inventories across the region. GLEC publishes it as a surcharge on the leg's energy: 12% for trucks above 3.5 tonnes in Europe, South America, Asia and Africa, and 15% for smaller vans. Rail takes the same 12%, borrowed from road for lack of its own data — the document says so explicitly.

At sea there is no surcharge but separate factors for dry and reefer containers, expressed per TEU-kilometre, and the gap between them is roughly a factor of two. For air freight, GLEC publishes no surcharge at all.

Picking the biggest truck

GLEC's classes run from 335 g CO₂e per tonne-kilometre for a 3.5–7.5 tonne rigid down to 54 for an articulated up to 72 tonnes. A factor of six. When a system offers «truck > 32 t» as the convenient option it is offering the low end of the table: the class recorded has to be the one actually hired.

Ignoring the empty return

The standard calculates on a round-trip basis precisely so the unladen return stays inside. Multiplying laden kilometres by a table factor is fine, because that factor already assumes a share of empty running; building your own factor from outbound fuel consumption is not.

The TEU-to-tonne assumption

At sea the factor comes per container and has to be converted to tonnes. GLEC suggests ten tonnes per TEU as typical, but a forty-foot reefer of fruit carries considerably more, and that assumption moves the result more than the choice of factor does. It is derivable from your own records: tonnes shipped over TEUs contracted.

Mixing methods inside one category

If inbound freight is calculated on one basis and outbound on another, category 4 ends up with two methodologies inside it and its total stops meaning anything. A declared difference between two reports beats a silent mix inside a single category.

• Classify every route by who pays before hunting for a factor: that fixes the category, and on exports it follows from the incoterm.
• Cut the chain into legs at each change of vehicle and each hub, and give every leg its mass, distance and mode.
• Declare the boundary — full or minimum — and hold to the same one across every route in the inventory.
• Ask carriers for fuel consumption on the routes that carry the most weight. For the rest, the table is enough to know where to look.
• Flag which legs are temperature controlled, and chase the refrigerant recharge data separately.
• Check the vehicle class recorded on each route before comparing years: that is where reductions that never happened tend to hide.